[Post updated -- see note at bottom]
How? They have now added search functionality.
By going to Archive.recapthelaw.org, you can search all the documents gathered by the RECAP Firefox Extension. The simple search allows you quick access to documents from U.S. Federal District and Bankruptcy Court documents, without charge. Further, the search will pull up the full docket sheet for a case, alerting and allowing you to acquire documents that are on PACER (and not yet in RECAP).
There is a simple and an advanced query page. The advanced query allows you to search by case name, number, court and dates of filing.
Two great features: it provides an RSS feed and an e-mail alert option for your query if you want to track the case.
Further, once you are viewing a case, you can add tags and connect related cases. For the case that I was searching, it even showed that it had been viewed 2 other times (by me).
As to privacy concerns, when you view case details, there is a button for reporting privacy violations. Although we hope that these privacy mistakes won’t occur, this button does help remove that data from the searchable RECAP archive.
The site mentions that it is still in the experimental phase and they welcome feedback. I might suggest a page of search tips (for example, should we use quotation marks for exact searches? etc). Also, I am not sure if this is searching full-text through all the downloaded documents and the tags that users might supply. Or, is this searching just through a few fields in the documents (attorney name, nature of suit, etc)? So, a bit more information on the searching will be helpful. [I will find out more and add to this post with details.]
But, all that being said, what a great resource. Now, what should we ask the folks at CITP to do for us next?
[UPDATE: this comes from Dhruv Kapadia, one of the Archive.Recapthelaw.org developers: "Right now the search is limited to the contents of the docket - the descriptions of each document as well as some parts of the metadata
associated with the docket. In the future, we may try to incorporate the OCRed text of the documents themselves, but we aren't doing that currently."]